deveation question 2

[beebo666]

is it acceptable practice to show two negatives in a tolerance. for example .756 -.001/-.002? thank you.

 

[ewh]

I have seen this done on manual drawings, but it doesn't meet the requirements of associative dimensioning with most CAD drawings/models. The model should represent the actual part, and such a dimension would not reflect the size of the actual hole without manually editing the dimension.
While I am unaware of any, there may be CAD packages out there that will handle this type of dimensioning.

 

[ajack1]

There certainly are CAD systems that can work this way and it is a perfectly normal way to dimension shafts and holes.

 

[aardvarkdw]

Take a look at your tolerance stackups, there may be a good reason for this.

David

 

[ewh]

ajack,
I was hoping that there was some software that could handle this (model at actual, dimension at two negative or positive limits - not actual).

 

[ajack1]

There is. Vero certainly does it.

 

[aardvarkdw]

Inventor will allow you to do this!

David

 

[ringman]

Dumb question, but "Why would you want 2 minus tolerances?" Have not seen or heard of.

 

[aardvarkdw]

Nominal size hole with a press fit maybe?

David

 

[beebo666]

man i really opend a can o worms here. but i asked because i got a customer drawing with that double minus tolerance.and from what i have read that is acepatable in metric drawings but not is US units. thanks for all the responses.

 

[ajack1]

It is certainly an ISO standard.

http://www.roymech.co.uk/Useful_Tables/ISO_Tolerances/ISO_LIMITS.htm

 

[drawoh]

ewh,

I did not know there was a CAD program that did not support double (-) or double (+) tolerances. The advantage of this procedure is that your dimension shows the nominal dimension. This is useful design information, however weird it looks to the fabricator. What matters to you?

I suppose you need a convention in your CAD office about how you model dimensions. I model to exact nominal size, and I apply tolerances. If I have to design something that mates to your part, I must examine your drawings, or at least look at the tolerances you applied to your 3D_model. This is true even if I understand how you scale your features. The size of your model, however you did it, is not an adequate guide.

JHG

 

[ewh]

We have to model to the actual finished dimension. We could use double negative or double positive tolerancing, but we would either have to lose the associativity to the feature by editing the dimension, or model the feature at the nominal, which may not reflect the actual finished part.
Please don't get me started on non-associative dimensioning. They cause nothing but problems for us.

 

[aardvarkdw]

ewh,

What program do you use?


David

 

[ewh]

UGS NX4

 

[ctopher]

I also don't know why anyone would do this. I have never seen it. But, doesn't mean it isn't possible.

Chris
SolidWorks 06 5.1/PDMWorks 06
AutoCAD 06
ctopher's home (updated 01-18-07)

 

[ewh]

I understand the reasoning behind its use, but I think it may be becoming rarer as MBD comes into its own. Depends on how the software is used and what the company dictates. To use it effectively, you have to model to nominal dimensions rather than actual, which may cause problems downstream in regards to any further analysis on that model or any higher assemblies.

 

[aardvarkdw]

I have never used UGS so I can't say anything in respect to it but I know Inventor allows you to dimension nominally and then appl a tolerance directly to the model dimension which is then parametric between the drawing and the model. I beleive Solidworks has this cabability as well. I agree with Chris though I don't see why you would want to do this except under a pretty specific set of circumstances. One instance that comes to mind is one in which you have a nominal size part like a dowel pin that you are then polishing to a slip fit. The nominal dimension would then be larger than any acceptable tolerance. Just a thought...

David

 

[ewh]

David,
When you dimension nominally, is your part modeled at nominal?

 

[aardvarkdw]

Yeah, which means that you don't have a graphical representation of the tolerance and if you use the measure tool to check the feature size you get the nominal size. The sketch dimensions show the tolerance if you tell the program to display them that way.

David