*Simple* Shelling Error with "Verification on Rebuild" Enabled

I guess I do not follow directions well because I can not get it to fail..

I think we ought to be able to shell something like you described, but I also can understand why it is difficult for the modeler. I agree that its annoying that SW can find a solution when it doesn't execute the thorough error checking of a verification on rebuild.

I vote bug, though faces that ram into each other are going to be a problem until they do a complete overhaul of their shelling methodology. For the folsk who want to try this at home, none of the dims matter - all you need is a wall section at the bottom of a pit that is the same as the shell thickness.

nevermind, I see now and posted it without verification on.

Yah know, if I used verification after rebuild "on" I would never get anything done. If this is a Parasolid issue, this may never be resolved?

..

Thanks Ed and Paul for verifying. Paul's model was exactly what I described. I guess I'll submit this finding and see what happens.

Cheers, Mike

Kind of like the previews that work on lofts and sweeps, followed by much swearing when it won't do it for real. Sometimes I can figure out what I screwed up. Other times I just have to start over again.

I vote for bug as well. It works in earlier code.

Jerry Steiger Tripod Data Systems

Eeek - I don't know if I have your courage. I agree that it takes time, but 'verification' finds problems, early. When I used to work with it off, I would get mysterious cherries popping up in features that built correctly before. Of course, it turns out the errors were always there, just not discovered. And lets not talk about errors that show with upgrades in SW - with verification on rebuild always running, I strongly suspect that we get fewer critical problems when we open parts in a new rev of SW. (But I have no data to support that - its just a hunch).

Both perspectives are valid, I suppose, but I prefer to work on a model that has been checked and is error free. I can always work around any problems that come up (at least so far I have...)

Just another perspective

-Ed

Just a little comment from a mold designer.......

I get SWX models in from people for new jobs and the first thing I do is try to scale it for a shrinkage factor. Occasionally, it won't scale (I get an error message about bad faces.) and I need to call the customer and tell them that I cannot use their model.

Many times in the course of the conversation I just described, I will ask if they had `verification on rebuild' checked. They will always reply with variations of what Paul said about not being able to get anything done if they check it. I swear under my breath, and add a few hours to their quote to cover me for the time that I'll need to spend cleaning up their model so I can design their mold.

Just my 2 cents.........

jk

We design and manufacture rotational mouldings which are formed by cavity only tooling with the internal surface formed by the moulding process. Therefore the logical way to model most parts is to produce the outer shape as a solid and then shell it to form the inner.

SW shelling is in my opinion pretty good compared with other Cad systems but varies from release to release. For example 2004 handles some types of self intersection better than 2003, but fails more with other types, it also seems to be more sensitive to complex surfaces. We use this as a design check, if the Cad model won't shell there is a good chance that there will be a material flow problem in the moulding, but the complex surface issue with 2004 makes this less useful than it used to be. If you have the freedom to change the design slightly, adjusting the geometry of the model where the failure occured is the only quick work around we have found. If you cannot change the design then surface modelling the part and then thickening it may be the only option but can take a lot of time!

It could be worse though, for example during a demo of an early release of Catia version 5 (5.3?) I saw the following > model a cube, put a variable fillet on one edge from 5 mm - 15 mm, shell at 10 mm, hello desktop! With SW and straight or smooth transition of the fillet, no problem.

Regards,

Jon

John,

Interesting, I personally would not send a SW feature part. I will send a SW dump/static brep or X_T to any mold designer unless there is a heck of a lot of trust in that relationship. Otherwise, regardless, the problem with sending a SW featured part are the issues per service packs, possible parameter changes, topology changes and the issue raise here.

For your work, wouldn't it be safe to work with only with the brep or Insert/Part? That is, Insert/Part and scale the inserted body, not scale the original SW featured part or is that what you are doing?

..

I hear ya. I guess over the years when I test verification on I do not see a huge issue but I do see the loss of time when I have it on. So, I've waged it is not something I can tolerate using. And the data I send out is static X_T data, as I mentioned to John.

..

Paul,

I have not been inserting the part and scaling, but that's not a bad idea for some parts.

The reason I prefer to work with a SWX featured part is that there are many part designers out there who don't understand draft and shut off angles. ( I am not speaking to present company.) It is often handy to have access to the feature tree so that I can easily make the necessary revisions. I always send this back to them for approval, but if I had to wait for them to figure out what to do, we'd never get a mold built.

Very often when I would contact them and mention draft, they would get very frustrated with me and tell me that was MY job. Others are just grateful that I can take care of it for them. Another issue is that I often am removed by a couple of layers from the customer. Say I get a job from a mold-maker with a 6-week delivery. The clock is ticking when he gets the P.O. He got the job from a molder who got it from the customer. To send the model up that chain and expect a good result to come back to me doesn't often work. I make the changes, and send the model out for approval while I continue with the tool design.

Kind of getting off-topic here, but that's what often happens. I fully appreciate what you say about maintaining the integrity of the model and not mucking with somebody else's well-thought out product design. If that design includes proper draft angles (as yours do) then just getting an x_t file would be fine.

jk

John,

You need to get a better class of customers!

Jerry Steiger Tripod Data Systems

It's not so much MY customers, it's the people that my shops need to work with to keep their doors open.

jk

OK, here is the answer... This is simply a function of the better (more complete) body/face checking with verification on rebuild selected. At the exact value .125 the top "face" of the cute coincided exactly with the shell offset. This would occur even if there was not cut feature (ie if you created this solid with a revolve) you would still get this error).

In fact, if you save this file out as a parasolid, import it, and try to add the shell, you get a reasonable error message:

"The shell operation failed to complete. One of the faces may offset into an adjacent face, a small face may need to be eliminated, or one of the faces may have a radius of curvature which is smaller than the shell thickness. Please use Tools Check to find the minimum radius of curvature on appropriate faces. If possible, eliminate any unintended small faces or edges".

Sounds like a limitation to me.

Mike Wilson

Still sounds like a bug to me. Works just fine in SW01+, whether I extrude, rotate, or bring it in from Parasolid.

Jerry Steiger Tripod Data Systems

Well, imho, there should be 2 possible solutions with the planar boundary zeroing out.

..

"Mike J. Wils>