Dimensional Tolerancing

[soupy1957]

If you saw a "2.0" on a mechanical drawing, and wanted to attach a tolerance to that number, (assuming of course that your "Tolerancing Data Block" is like ours and doesn't typically include a tolerance for a ".0" number), what tolerance would you go to, by default? ±.010 (or higher)?
-Soupy1957

Telepathy only enhances the world I live in!!

 

[vc66]

You may get more responses if you post this Drafting Standards, GD&T & Tolerance Analysis Forum, but the answer to your question is simple.

How big of a tolerance can you functionally allow? Obviously the bigger the tolerance the easier to create the part, but you can only make the tolerance wide enough so as not to disturb form fit or function. There is no "standard" tolerance.

V

Mechanical Engineer
"When I am working on a problem, I do not think of beauty, but when I've finished, if the solution is not beautiful, I know it is wrong."

- R. Buckminster Fuller

 

[aspearin1]

By definition of significant figures your tolerancing cannot be any lower than +/- 0.1 in increments of tenths (2.0 +/- 0.1 or +/- 0.2 Not +/- 0.15). Why does no one use SigFigs anymore? Do the extra decimal places really cost too much ink on the drawing? From a non-drafting point of view, but as a process engineer, I would assume a tolerance somewhere between 0.1 and 0.3, if I had no other information I could go by.

Aaron A. Spearin
ASQ CSSBB
Engineering Six-S'$

http://www.Engineering6ss.com

"The only constant in life is change." -Bruce Lee

 

[vc66]

You're absolutely right aspearin1. I didn't include the fact that you couldn't have .01, .001, .0001, because I figured it was understood that you couldn't use them. Oversight on my part, as the OP used +/-.010. With one decimal place as your nominal, your tolerance can only have 1 decimal place, i.e. you can't have a dimension 5.0 with a tolerance of +/- .0005.

V

Mechanical Engineer
"When I am working on a problem, I do not think of beauty, but when I've finished, if the solution is not beautiful, I know it is wrong."

- R. Buckminster Fuller

 

[aspearin1]

Just a pet peev of mine, I suppose. I recently got an exam question wrong where the question was worded with 1 sigfig, and the exam asked for an answer with 5 sigfigs. Technically, there was no right answer.

Aaron A. Spearin
ASQ CSSBB
Engineering Six-S'$

http://www.Engineering6ss.com

"The only constant in life is change." -Bruce Lee

 

[eromlignod]

"Why does no one use SigFigs anymore?"

Because default drawing tolerances can give a more appropriate tolerance for the work.

Our drawings, for example, give a tolerance of +/-.005" for three-figure dimensions. This is a reasonable amount for general, non-NC machine work like tapped and clearance holes that have to line up between parts. The sig-fig tolerance would be .001", which is finer than necessary and is more difficult to produce by a machinist, especially if he is just scribing and center-punching by hand. And to show the dimension with two decimals would indicate a sig-fig tolerance of .010", which could mean that two holes could be off by as much as .020" relative to each other, which is a little too sloppy.

I think engineers give tolerances that are too tight more often than they do ones that are too loose, which can greatly increase the manufacturing cost of a part.

Don
Kansas City

 

[TomFin]

what tolerance would you go to, by default? ±.010 (or higher)?

The drawing block does not contain a specification for the tolerance range of a 0 .X callout. I agree that one has no other option than to logically resort to sig-figs. But realistically there is a narrow likelihood of an inspector or machinist of recalling sig fig rules with enough confidence to make a pass fail decision. I think the best solution would be to figure out how much slop is allowed.
If the drawing is not controlled by your organization then I would think you have as much tolerance as sig fig rules allow for. If you do have jurisdiction and influence over the drawing block standards then modify the drawing block to specify tolerances for such callouts.









Failure is a prerequisite of successful design

 

[Cockroach]

Higher. Typically single decimal point is +/- 0.03 inches.

Kenneth J Hueston, PEng
Principal
Sturni-Hueston Engineering Inc
Edmonton, Alberta Canada

 

[KENAT]

What does the tol block say/have?

If say the block tol is +-.010 for 2 decimal place and +-.005 for 3 decimal place and has nothing for 1 decimal place then I'd probably go to the +-.010.

At the end of the day though it's an incomplete drawing so you should probably ask whoever the drawing came from.

As to the sig fig stuff, unless somehow invoked by a note or similar on the drawing I'm not sure how you'd justify invoking it.

KENAT, probably the least qualified checker you'll ever meet...

 

[fcsuper]

Sometimes a signal place 1.x dim may already have a tol mentioned in the drawing notes. If not, I agree with KENAT about using the drawing's largest tol. However, it sounds like one way or the other, the drawing needs to be corrected to have a tol.

Matt
CAD Engineer/ECN Analyst
Silicon Valley, CA
sw.fcsuper.com
Co-moderator of Solidworks Yahoo! Group

 

[KENAT]

thread286-201572

KENAT, probably the least qualified checker you'll ever meet...

 

[jbwick]

Do you have a shop standard? Many of the drawings I have used and written have default language that refers to a shop standard for anything that is not explicit on the drawing. It is an incomplete drawing as someone else noted.

Additionally I could not even begin to guess an appropriate rule of thumb without knowing the material of the part and the process used to put it in spec.

a 2x4 cut to 2.0 feet with hand tools ... I would likely use +/- .25"

On the other hand a 2.0" long 1/2" diameter aluminum rod on a lathe +/- .005 is acheivable if nearby features call for it.

 

[KENAT]

jbwick but of course, it should be driven primarily by function, not what typical processes can achieve, no?

KENAT, probably the least qualified checker you'll ever meet...

 

[MechEng2005]

As already stated, the drawing is incomplete and any implied tolerance is guesswork... If I were trying to argue for/against a part with that dimension that failed/passed inspection I would assume that it would be +/-0.05

If it was produced to 2.06, then for one significant figure the dimension would be 2.1, which does not meet the 2.0 stated. Similarily, if the part was produced to 1.94, this would be 1.9 in one place form, which doesn't meet 2.0 dimesion.

-- MechEng2005

 

[jbwick]

Kenat I agree function should be the main concern. But there are other limiting factors including material availability, available processes and cost. Unfortunatly cost frequently has an unbalancing influence on these decisions.

Fot the OP my idea is to evaluate the needs of the part or the assembly it goes into and use the largest tolerance that will result in a working part. You are the only expert on the use of your part involved in this discussion. Every idea presented here will produce a part, will it work?