BoundingBox vs Create Box, incorrect results from BoundingBox

[BBow]

We have implemented and successfully been using the journal from:
thread561-389240
to run BOM finished sizes for our components for about a year now. Works perfectly 95% of the time. However there are certain instances where it was failed us, and they are seemingly repeatable. We have found that chamfering edges on 'broken out' holes/cutouts on a relatively thin plate causes the plate thickness to be called out as thicker than actual model (eg. a 1.00" plate with a cutout with a chamfer will display as 1.014" and this will change as the chamfer size changes, and then be exact when the chamfer is suppressed). I'll attach a picture. This is not the only feature that has caused an error, but the most prevalent and known off the top of my head.

1) Is there a fix or known reason for this?

2) Why does this BoundingBox function work differently than the CreateBox function? When we manually do a Create Box then measure it, it is always exact???

Any ideas? Like I said 95% of the time it works great and has still solved many issues and saved us a lot of time, but never hurts to try to improve something to 100% right?

 

[cowski]

From the API help files regarding the .AskBoundingBox function:

Note that the box returned favors speed over accuracy. The returned box will contain the given entity and it will usually be close to the minimum bounding box but this is not guaranteed.

There is a more accurate version available, called .AskBoundingBoxExact, that you might want to investigate.

www.nxjournaling.com

 

[BBow]

Thanks for the quick info, I'll check into it.
As always thanks!

 

[Toost]

Only to inform new readers , There is from, i think, NX10 a feature named "Bounding Body", which creates a enclosing block/ cylinder. - without running custom code. - Use command finder.

To compliment the thread, sometimes NX uses facets to calculate things ( Like internal volume around other solid bodies ), i would guess that the imprecise box you get with you code is calculated that way.
- and thereby the other .AskBoundingBoxExact function.


Regards, Tomas

 

[BBow]

Works like a champ, fixed all issues on known parts I've been able to test.

Thanks again