Flatness vs. Parallelism

[koooala]

Hi,

I would like to know if there is principle of tolerance or dependence between Flatness and Parallelism ?

for example:

cube with [datum A flatness 0.02 (mm)] on its one face
and on the parallel face tolerance of [parallelism 0.02 (mm) to datum A]

the value of parallelism tolerance should be bigger or smaller than flatness ?

Thanks

 

[cjccmc]

"the value of parallelism tolerance should be bigger or smaller than flatness ?"

Assuming you mean bigger or smaller relative to the flatness of datum A surface, I think it can be either, depending on what you need. You could also call a value for each , for example, parallel .04 A, flat .01.

 

[SeasonLee]

When parallelism is applied to a surface, the tolerance zone is two parallel planes which are parallel to the datum, all elements of the tolerance surface must be within the tolerance zone, the flatness of the surface is also controlled within the parallelism tolerance zone.

SeasonLee

 

[CheckerHater]

It's kinda like this:

http://files.engineering.com/getfil...bf33-484c-ac71-fd99ff92ad94&file=FLATNESS.bmp

 

[pmarc]

koooala,
From your descripition it seems that you need to know whether there is any Flatness / Parallelism relationship when those two characteristics are applied to different features. If that is the case, the answer is NO there is no such relationship, however...

If you imagine your cube having a linear dimension specified between the two sides, then the values of these geometrical callouts will be dependent on dimension (size) tolerance. That being said, if the dimension was 10+/-0.01 (i.e. 0.02 tolerance), both values you gave would be acceptable, but could never be greater than 0.02. This is assuming you work to Y14.5 when envelope principle is default rule for tolerancing.

 

[koooala]

CheckerHater:" It's kinda like this: -> [link

http://files.engineering.com/getfile.aspx?folder=e...

]

http://files.engineering.com/getfile.aspx?folder=e...

[/url] "

It is exactly what I had in my mind.

Does every body agree with this picture ?

summary the picture:
1. Parallelism tol shouldn't be a smaller than Flatness tol
2. Size tol shouldn't be a smaller than Parallelism tol

What should be the recommended proportion of this values ?
i.e Size Tol = (Flatness Tol)/10
or Size Tol >= Flatness Tol
or Size Tol >> Flatness Tol

or Parallelism Tol >= (Flatness Tol)/2

Thanks

 

[dgallup]

Generally speaking, I don't see any point in a geometric tolerance that is greater than half the size tolerance. Beyond that it just comes down to what you need. There are no hard and fast rules.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[CheckerHater]

koooala.

There is no rule chiseled in stone, but you may consider the following ratios between size tolerance and associated form tolerance (flatness or straightness) in relation to level of precision:

Normal (or “standard” or “medium”): 60%
Enhanced (or “reduced tolerance”, or “fine”): 40%
High (or “precision” or “extra fine”): 25%

Once again, it’s not a standard, just a general guidance.

 

[KENAT]

Hmm, I'd be very careful using rules of thumb for this.

We make precision equipment which for several reasons often need very good flatness and/or parallelism.

It's not unusual for us to have parts with say +-.010 size tolerance on a dimension but with flatness of under .0005 and parallelism of around .001 or similar.

I'd say this is a case where it really should be driven by functional requirements.

Posting guidelines faq731-376

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

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

 

[CheckerHater]

Kenat,

I was just trying to illustrate what could be considered “normal” and what “precise” from average shop point of view, not “your flatness should always be ¼ of your size tolerance”.
You said it yourself: “We make precision equipment”; so I assume sometimes special considerations are taken to achieve required flatness, etc.

Could we add “or less” after my numbers, pleeease?

 

[fsincox]

CH,
I always wondered if there might be some kind of rule of thumb for this kind of thing. I do understand it is not a hard rule!
I am curious where it come from, I always thought it might make sense to be like 1/3 of the total size (rule of Frank);]

 

[fsincox]

Ken,
I agree, I am just interested in the subject and functional requirements are not always so clear cut.
I think this is also a "machined features" type of thing, CH, correct?

 

[CheckerHater]

Frank,

This is from the book printed long ago and far away. There is small possibility to find old ISO standard on this topic.

And I agree: sometimes “fairly flat machined surface” IS functional requirement

 

[pmarc]

I agree that the values for straightness or flatness tolerances shall be chosen based on functional analysis, however in order to pick appropriate values one must understand how accurate a manufacturing process is - in other words, which values are realistic and achievable, and which are not.

This is what ISO 2786-2:1989 says about general flatness and straightness tolerances for machined part. Interesting is the fact that the values are not dependent on the thickness of the feature but on its length. Another interesting thing is that those values are named as corresponding to "customary workshop accuracy".

http://files.engineering.com/getfil...126988&file=gen_straightness_flatness_ISO.png

 

[dgallup]

The flatness or parallelism geometric tolerance must always be a refinement of the general tolerance of size. So it makes no sense for it to be equal or nearly equal to the size tolerance. That is why I said as a general rule is should be about half or LESS. It can be as small as you need it to be. I've got flatness tolerances that are less than 1% of the size tolerance where a part has to be really flat. It takes some specialized manufacturing to do it.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[KENAT]

I'd lean toward what dgallup is saying.

If envelope principle is in play then your flatness/parallelism tolerance at worst case = you size tolerance. Now in practice you're going to get significantly less than this in most cases as it needs to cover both flatness, parallelism and other surface deviations.

Hence if you need a separate flatness/parallelism tolerance beyond what you get from envelope principle then it generally makes sense that it's going to be less than around half your size tolerance.

Pmarc, while manufacturability is important, it should be secondary to function. You need to make sure what you functionally neeed can be made cost effectively.

However, just because flatter than you need is typical of the anticipated manufacturing process, I'd be loathed to specify a tighter flatness than I actually need. Processes change, small errors get made... so I'd leave my flatness as loose as functionally permissible.

Posting guidelines faq731-376

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

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

 

[dgallup]

Kenat - that's twice you've agreed with me today. Either you're getting soft or my posting is improving.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.