Anyone ever see a double positive tolerance? 6

[jaimeshawn]

Anyone ever see a double positive tolerance? (e.g 1.000 +0.005 +0.008)

I've always thought the target value must be between the maximum and minumum value; and was selected by the engineer based on maximizing yield depending on how the part would be made, and on maximizing performance of the final assembly. The tolerances could be symmetric +/- 0.003, or they could be asymmetric +0.003 - 0.000, but I've never seen double positive tolerance, or double negative before.

Did I miss something in GD&T class?

 

[handleman]

29.95 is not nearly as nice as 30. 29.9865±0.0135 is downright hideous compared with 30g6.

-handleman, CSWP (The new, easy test)

 

[handleman]

Oops, should have been 29.9865±0.0065

-handleman, CSWP (The new, easy test)

 

[KENAT]

Depends who's using it.

I know plenty that 30g6 would be meaningless to, especially without the relevant ISO referenced on the drawing.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

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

 

[handleman]

Exactly my point (well one of my points anyway Dang shotgun brain), KENAT - depends on who's using it.

I do usually include the -0.007/-0.020 values along with the tolerance class.

Also, from what I understand, the machinist will not be aiming at 29.9865 anyway. If it's a shaft, he'll shoot for just under MMC. Lots easier to fix if he misses. It's easier to get to 29.993 by subtracting 0.007 from 30 than by adding 0.0065 to 29.9865.

-handleman, CSWP (The new, easy test)

 

[TheTick]

I also don't like the idea of creating a feature in CAD with a size that is outside of its tolerance band. As was discussed above, it can cause problems when cutting parts directly from CAD data. Sure, the programmer should always double check, but "should" is a word that can waste a lot of time and money if used too liberally.

 

[dgallup]

Quote: 29.95 is not nearly as nice as 30. 29.9865±0.0135 is downright hideous compared with 30g6.

If you put 29.9865±0.0135 on the drawing you have to have a gage that is accurate to .00005 which is pretty near impossible. Any machinist that is not familiar with the ISO system of fits should not be making your parts. Normally parts requiring these levels of precision are not going to be made by the illiterate. If they are being made in volume there are probably automatic gaging systems in place with nice green & red lights that anyone can understand.

However, if you don't need asymmetric tolerances, definitely don't use them.

 

[KENAT]

"Any machinist that is not familiar with the ISO system of fits should not be making your parts. "

Having picked my self up from the floor where I'd fallen laughing. For those of us not caught up in the big defense & automotive supply chains, but too big to have much say in who the parts get sent to, I'd be amazed if Purchasing/QA even considered that issue.

Even in the UK we rarely used the "30g6" form, we'd put down the equivalent dimensions. Only exception I can think of is for threads.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

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

 

[ctopher]

A tool to help with future tolerances. Browse around the site, there are more.

http://www.engineersedge.com/true_position.htm

Chris
SolidWorks/PDMWorks 08 3.1
AutoCAD 08
ctopher's home (updated Aug 5, 2008)
ctopher's blog
SolidWorks Legion

 

[DanStro]

I've used them quite a bit but I've since replaced them with limits. Mainly because I got tired of machinists either calling asking if this is a typo and do I really mean +xx/-xx or, even worse, assuming that it is a typo and I get bad parts. It just wasn't worth the hassle.

 

[dingy2]

ctopher:

I have a bit of a problem with the web site you suggested when positional tolerances are called "true position". ASME Y14.5M-94 states that per 1.3.36 on page 4 True Position means "The theoretically exact location of a feature established by basic dimensions." The term used should have been "positional".

I might be nit-picking but a suggested web page and training advertisement should use the correct terms to gain credibility.

Dave D.

http://www.qmsi.ca

 

[ctopher]

Dave,

I don't back any site personally or professionally, I just suggest some sites for reference occasionally. Engineers should 'always' refer to ASME Y14.5-94.

I don't always have the ASME spec in front of me because I'm either virtual or at another site. So, I usually just do a quick Google search and share some sites I think may help a little.

I agree with you they should use the correct terms. But unfortunately I sometimes don't have the time to peruse the sites first.

Chris
SolidWorks/PDMWorks 08 3.1
AutoCAD 08
ctopher's home (updated Aug 5, 2008)
ctopher's blog
SolidWorks Legion

 

[CheckerRon]

Off topic for this post, but "True Position" suggests your dealing with an old guy. i.e. thinking pre 1974.

On topic: ++ or-- tolerances are fine for fit tables, but I would never use them on a drawing. I either convert to limit dimensions or mean value with ± tolerance.

 

[ajack1]

I totally agree with handleman here.

“Check a catalogue for dimensional drawings of anything - linear shafting, bushings, servo flanges, spacer blocks, dowel pins, raw steel dimensions... the list could go on and on. If there's any sort of fit involved, you will see unilateral tolerances. I'm constantly amazed at people's ignorance of this common practice.”

The whole or should that be hole reason behind this is to standardise things so parts are readily available and interchangeable, if you accept that all these items and more have unilateral tolerances on them why should the parts that they fit not have as well? The other option is of course to make everything specials and increase your costs by about 400%.

If you work in a part of the world that does not understand what 30g6 means then by all means put 30 -0.007/ -0.020 but to put 29.9865 +/- 0.0135 is just plain madness as you are just increasing the cost for producing the part with no gain what so ever.

I to am constantly amazed at people’s ignorance of this common practice, I am even more amazed that companies that are forced to increase their costs through not understanding them still remain in business, they wouldn’t be here next month in automotive.

 

[handleman]

Hey, thanks for fixing my spelling of catalog(ue)...I fear I am a poor user of the Queen's English. ;-)

-handleman, CSWP (The new, easy test)

 

[CheckerRon]

Writing the peasants english, "catalog" is just fine.

I'm contracted to a company where the division DRM disallows unilateral dimensions on drawings.

We must use "target" (i.e. mean) dimensions with ± tolerances; equal, or unequal, but no + or - 0.
It is a shop, inspection driven requirement, much to the dismay of engineering.

Zero position tolerance an exception, of course.

 

[KENAT]

"If you work in a part of the world that does not understand what 30g6 means "

ajack, in the UK we had people that couldn't handle this so used to normally put limits.

KENAT,

Have you reminded yourself of faq731-376 recently, or taken a look at posting policies:

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