Model vs Drawing tolerances (Y14.41 ?) & notes.

[PTMALM]

We are using models as the "master data", yet have drawings to clarify what is not on the model. My issue is with tolerances. Typically untoleranced dimesions on the drawing default to the title block tolerances (.xx, .xxx). However, when using the model as the data, how are the tolerances handled. I had placed a note saying that any model measurement shall be considered a two place decimal per the drawing title block. I also have a note that states that the (hard) dimensions on the drawing superceed the model data, so that there is no mis-interpretation between the model and the drawing. Is this the correct way to pursue this? Since I have up to now provided complete details on the drawing I did not have to worry about this. Where I'm at now, they have been mostly involved with prototyping, and have been able to get by without this kind of control. But they are now trying to apply the same method to high production, which makes me nervous.

 

[MadMango]

I think you need to invest in ASME Y14.41-2003, Digital Product Definition Data Practices.

This defines how GD&T are applied to 3D models.

[green]"Art without engineering is dreaming; Engineering without art is calculating."[/green]
Steven K. Roberts, Technomad
Have you read faq731-376 to make the best use of Eng-Tips Forums?

 

[PTMALM]

I'm planning on ordering it. However, the customer is reluctant in even using GD&T (the barbarian factor) or I'd use a surface profile tolerance on all untoleranced values. But I'm still going to give it a try...

 

[ctopher]

I use the tolerance within the model the same as what is on the dwg. The problem is when someone else changes the tol on the dwg and not the model.
I do my best to make all dim's 3 place decimal nom, then add tol where needed.

Chris
Systems Analyst, I.S.
SolidWorks 06 4.1/PDMWorks 06
AutoCAD 06
ctopher's home (updated 06-21-06)

 

[MadMango]

If your customer is providing the models and the drawings, they should determine which is Law, not you. You would hope that the models are created at the mid-range of all tolerances, but we all know that hardly ever happens.

[green]"Art without engineering is dreaming; Engineering without art is calculating."[/green]
Steven K. Roberts, Technomad
Have you read faq731-376 to make the best use of Eng-Tips Forums?

 

[ajack1]

You have to think differently about dimensioning a model to 2D drawings. Say you have a surface made up of 300 parts how can you possibly dimension them all in 2D, especially if most and not rads or planes?

You have to use surface and trimline tolerance and just add positional tolerance to key features like hole positions.

 

[ewh]

I think that there would be less of a chance of misinterpretation if you made the model the master definition instead of letting on hard drawing dimensions supercede the model.

 

[drawoh]

PTMALM,

This is the same problem you get when you have fabrication drawings with both metric and English dimensions. You have two sources of information, one of which is correct, and one of which you provide entirely out of the goodness of your heart, for reference only. There is no other way, other than not providing the second set of information.

Your master data must include a complete set of dimensions and tolerances, somehow. If you cannot implement ASME Y14.41-2003, then you must prepare complete 2D drawings, and provide the 3D model for reference.

I used to work as an inspector in the automotic parts industry. I never saw a drawing. As far as I can recall, the customers provided inspection fixtures. Maybe this is your solution.

JHG

 

[PTMALM]

ewh, the issue with the drawing superceeding the model would be (for example) in the case where the part must have a tighter tolerance than standard, or must be a plus zero, minus something for example. Since I am not allowed to go "all model" and utilize the y14.41 capabilities of SW, I needed some way to "over-ride" the model.

 

[ajack1]

PTMALM that is exactly how all the major automotive manufacturers work for panels. The model will have a general trim and surface tolerance on it and certain areas will be shown as say a unilateral tolerance and most holes and slots unless for weight reduction or assembly access will have a positional and size tolerance unique to them.

It is also common to show the areas where the part should be clamped for inspection and the datums.

 

[AnnaWood]

As ajack1 says this is how automotive does it and has done it for a long time. We were creating minimum content drawings with the math data as the master back in 1998 when I last worked for the GM Truck Group. Doing details for body panels on maybe a dozen 11 x 17 sheets.

I can't wait for the day that the business I am in now, tool and die, can fully utilize minimum contest drawings. We have all the equipment to get'er done that way. Only thing standing in the way is the mindset of the people who only know one way of getting things done.

Regards,

Anna Wood
A former automotive sheet metal and interior trim senior designer, CGS jockey, Philpot and DeSmet grad.

 

[ewh]

PTMALM,
This is often the purpose of reduced dimension drawings. Only show the dimensions that are critical and must be held to a tolerance other than the general tolerance. Not everything needs to be dimensioned.

 

[ajack1]

Anna many tool and die manufactures in the UK do work to solid models with little or no dimensioning, or at least the smarter ones do, especially on bigger tooling with a lot of castings in it.

As with any “new” technology some people grasp the concept and others play at it. To me at least to start creating 2D drawings from a model is like constructing a profile in CAD and then dimensioning it up so someone else can create the same profile, or drawing something in 2D printing it off and dimensioning it on a drawing board, you simply take away the main advantage of doing it that way in the first place.

The main advantage of modelling as I see it is poly patterns or castings can be made to them, simulations can be run on them, machining can be done to them, inspection can be done to them, they are the master for everything. I think many companies that fail to realise this will fall on hard times.

 

[KENAT]

Are the models in question smart or dumb?

If you are talking dumb (step, iges etc) models surely you need at least a reasonable drawing for tolerances, material, notes etc?

If the customer/supplier is able to read annotated models then I can see how MBD may work but I'm still struggling to get my head round it.

Maybe if I ever manage to get 14.41 it will all become clear.

 

[fcsuper]

Under ISO, you shouldn't even have an issue, since a feature's size determines its tolerance. However I work in an ASME enviroment. We have this boilerplate note: "THIS IS A CRITICAL TO FUNCTION DRAWING ONLY. IT IS TO BE USED TOGETHER WITH THE CORRESPONDING CAD FILE TO FULLY DEFINE PARTS. ALL DIMENSIONS NOT SHOWN SHALL HAVE TOLERANCE OF +/-.0XX"

Our parts are in the size range where it's not important to define different tol's for different size features, so this general note works for us and we don't even need ISO-like conventions.

If you deal with parts/features of drastically varying ranges, then perhaps emulating the ISO tol convention might be advisable for creating standard tolerance on your solid models.

Matt

 

[KENAT]

My last place had a very simple block tol, just a single figure for linear and single figure for angle. It depended on neither size or decimal places. If the dimension didn't have a +- or limits or other tol then you applied the block tol.

The designer selected block tol based on what he thought was going to be the most common/applicable/required tol on the drawing.

Any dimension needing a different tol was modified to be +- or mor normally limits.

Bad thing was you had to look at every dimension and think what tolerance was applicable.

Good thing was you had to look at every dimension and think what tolerance was applicable.

 

[KENAT]

Sorry should have been single value. The value could have had more than one figure. For metric we usually used 0.25 for inch usually .01. We would adjust this though eg welded assemblies usually had larger tolerances if dimensioned etc.