Flatness & Surface Finish

[Smile9999]

for example, if I have a plate that needs to have a finished and flat surface, what would make me choose surface finish over flatness or otherwise?

I suppose my question is really about why would I ask the machining to make me a flat surface or a finished surface when the flat surface would do for both flatness and surface finish?

btw I am talking from Geometrical Tolerance perspective, where I can just put a flatness tolerance instead of a surface finish symbol

 

[JohnRBaker]

'Flatness' is an indication of how much 'warpage' or deviation from being planar is a surface that you're willing to accept. While 'finish' is how 'smooth' (or how 'rough') a surface is. You can have a very SMOOTH surface which is NOT flat, while you can also have a very FLAT surface which could have a ROUGH finish. The two are not really related as they measure totally different characteristics of a surface.

John R. Baker, P.E.
Product 'Evangelist'
Product Engineering Software
Siemens PLM Software Inc.
Industry Sector
Cypress, CA
Siemens PLM:
UG/NX Museum:

To an Engineer, the glass is twice as big as it needs to be.

 

[Engdoitbetter]

I would add that, while in practice tolerances and roughness are treated as two different topics, in reality there is a connection between them.
In fact, when measuring roughness, you first have to "read" the surface geometry with the proper instrumentation, then you must filter the data so that any macroscopic deviation from the ideal shape (i.e. geometrical tolerance) is eliminated. What is left is the micro-geometry of the surface, which we call roughness. Then, though tolerances and roughness may not be generated by the same phenomena, from a geometrical perspective they are the same thing, but of different order of magnitude.

Leaving behind academia, geometrical tolerance controls how much a particular shape is allowed to deviate from an ideal one (e.g. a pin must be cylindrical enough to rotate well inside a hole), while roughness measures the finish of a surface (e.g. specular, as-welded, as-cast etc.), which is, for example, important in seals (different types of gaskets require different types of finish).

Hope it helps.

Stefano

 

[mfgenggear]

It is standard practice to call a flatness & surface finish sperately "if it's required", on a print. as separate entities.
roughness even thou measures max hollows & peaks & valleys, it is at a smaller measuring area.

It is standard for a machinist & inspector to verify each separately.
roughness symbols are generally not used to control flatness.

Mfgenggear

 

[jcrags]

Also, think about how flatness and surface finish are measured. You could measure flatness by dragging the part around under a dial indicator, and have the face seem perfectly flat, because the large surface area of the indicator spanned the small irregularities. But, that surface could have been milled in a way as to leave deep gouges. This is something that I deal with all the time with vacuum sealing orings. Just because a surface measures "flat" doesn't mean it'll hold a vacuum. Often, I end up having to pull out the sandpaper to improve the surface finish where the oring needs to seal.

 

[IRstuff]

From a 10,000 ft perspective, flatness is a low-frequency specification, while roughness is a high-frequency specification. That is to say, roughness deals with small features, on the order of angstroms and nanometers in diameter, while flatness deals with entire surfaces, i.e., inches, etc.

Therefore, while there might a few applications where flatness and roughness have the same numerical values, in general, flatness is rarely as small as the roughness requirement. In the case of optical surfaces, such as mirrors, flatness and finish might be comparable requirements, as any deviation from the ideal might create optical errors. Even then, since flatness is derived from wavefront error, while finish is derived from scatter, they still might be different, since they have different reasons for existence.

One of the downsides of not doing systems engineering on derived requirements is that the recipient engineer will often not know why a requirement is the way it is, since there are often disparate requirements that dictate the flatness and finish.

TTFN
faq731-376

 

[KENAT]

Please don't double post.

thread281-326455

Or if you absolutely feel you need the input of folk in more than one forum then have one 'main' thread and in any duplicates just put a link to the main one and ask respondents to post there - though even this sometimes upsets site management.

Posting guidelines faq731-376

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

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