General note for MBD Basic Dimensions

[dtmbiz]

Looking for input as to which of the two notes below show the most promise for Basic Dimension measurements in Model Based Definition files


1 ALL UNSPECIFIED DIMENSIONS SHALL BE QUERIED FROM THE MODEL AND CONSIDERED BASIC.
QUERIED DIMENSIONS MUST BE TAKEN BETWEEN THE MEASURED SURFACE AND THE APPLICABLE
DATUM REFERENCE FRAME. QUERIED DIMENSIONS SHALL BE ROUNDED TO THE SAME NUMBER
OF DECIMAL PLACES AS THE APPLICABLE FEATURE CONTROL FRAME.

or

2 ALL QUERIED DIMENSIONS SHALL BE CONSIDERED AS BASIC WITH A “MEAN VALUE” UOS.
DECIMAL PLACES SHALL USE THE SAME NUMBER OF PLACES AS THE APPLICABLE GEOMETRIC CONTROL TOLERANCE.
MEASUREMENTS SHALL BE TAKEN FROM THE RELATED DATUM FEATURE TO THE MEASURED FEATURE.
UNIT OF MEASURE SHALL BE THE SAME AS THE MASTER MODEL DEFAULT UOS.

 

[drawoh]

dtmbiz,

What does UOS mean? TLAs are a pain in the butt. The word "QUERIED" does not work for me. I don't see why anybody cares about decimal places. Dimensions are basic, not the model.

Tell them that unless otherwise specified, all unspecified dimensions are to be measured from the model. Give them a profile tolerance. Specify datums on the drawing.

--
JHG

 

[dtmbiz]

drawoh

UOS = unless otherwise specified

From what I recall (few years back); mean value, number of places, units, are all requirements from ASME Y14.41 Digital Product Definition. Now I hear that ASME Y14.100 is that standard (not available to me though)

Do you have a preference with #1 or #2 as a starting point for further development?


#1 concerns me for a number of reasons.
- Measure from the Datum Reference Frame could be interpreted as from the DRF origin. This could
cause a wrong measurement.
- unspecified ? what measurement is specified?
- Lacks information mentioned above.
- not all measurements would involve a surface

Thank you for your reply

 

[drawoh]

dtmbiz,

I don't like either of your notes. When I re-read my note, I find my use of the word "unspecified" to be unnecessary as well. ASME[ ]Y14.5M-2009 states that decimal accuracy is infinite. If your maximum size is called up as 6.0mm, 6.000000000000000000000001mm is too big. Decimal places are only relevant if you specify dimensions on the drawing, and rely on the tolerance note on your title block.

When you specify datum features on your drawing, you are telling everyone how to fixture the part for fabrication and inspection. Datums are way more than a dimension zero point. Section[ ]4 of ASME[ ]Y14.5-2009 explains datums in way more detail than I am willing to go into here. Datum reference frames are not a problem unless somebody does not understand the standard.

If you want to communicate meaningfully with people, avoid acronyms and initialism. I am a bit of an asshole about this because I can touch type, and it is no trouble for me to spell this stuff out. I assume you are creating a title block or some other form. You only have to do this once. Do it properly.

You need to state that, unless otherwise specified, all your dimensions are read from the 3D[ ]model. You need to provide an all-over profile tolerance (ASME[ ]Y14.5-2009 Section[ ]8.3.1.6). You can make your profile with-respect-to datums[ ]A, B, and[ ]C, but you need to be able to edit this conveniently because this only works 95% of the time.

--
JHG

 

[3DDave]

Y14.100 is the overall specification. It calls for 14.5, et al.

From the ASME page:

This Standard establishes the essential requirements and reference documents applicable to the preparation and revision of manual or computer generated engineering drawings and associated lists unless tailored by a specialty Standard. It is essential that this Standard be used in close conjunction with ASME Y14.24, ASME Y14.34, ASME Y14.35M, and ASME Y14.41. Incorporated into Y14.100-2013 is Y14.42 on Digital Approval Systems.

The problem with digital models is that the underlying number system is base 2, which means that few decimal numbers are accurately reproduced. It's necessary to determine how to truncate/round the base 2 values to get base 10 values. See "What Every Computer Scientist Should Know About Floating-Point Arithmetic"

https://docs.oracle.com/cd/E19957-01/806-3568/ncg_goldberg.html

or the short version

http://floating-point-gui.de/

 

[KENAT]

thread1103-411315

I there may be other threads that touch on this.

Posting guidelines faq731-376

http://eng-tips.com/market.cfm?

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​