roughness in flatness

[bxbzq]

Hello,
I assign a flatness error 0.1 to a surface and I get a part having surface profile shown in attached draft. From the definition of flatness in Y14.5-2009:
Flatness is the condition of a surface or derived median plane having all elements in one plane.
It means to me that the flatness error is A-C. But during inspection, I think the measurement equipment gives A-B. The difference B-C is from the roughness of the surface.

Any comments?

 

[Belanger]

I will presume that A is the bottom blue line, B is the middle red line, and C is the top line. Then, the distance A-B is the roughness, and the distance A-C is the flatness (but depending on the values given, A-C could also be the "waviness").

You'd have to look at the settings on the measurement equipment -- there are algorithms such as Ra and Rz for roughness, but those shouldn't be used if your only specification is flatness of 0.1.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[pmarc]

Why are you saying that the distance between the middle and the bottom line is surface roughness in this case?

 

[Belanger]

Again, we don't know the exact numbers or scale of things, so he'd have to talk to the inspector for more information before we lay down a definitive answer of what's going on. I guess that is the caveat in my initial reply.

Surface roughness is typically considered to be the high-frequency, short-wavelength component of a measured surface (that would be A-B). The larger-scale, general up-and-down pattern that you see (A-C) is usually considered as the waviness.

How were you looking at it?

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[bxbzq]

To avoid confusion, see attached. I think roughness value is close to distance B-C.
I want to confirm my understanding, that is:
By definition, the flatness error is distance A-C
In reality, the measured flatness is distance A-B

Thanks.

 

[IRstuff]

According to ASME B46 Surface Texture, roughness is the high frequency variation that rides on the waviness, so B-C is the peak to valley roughness, and A-B is the peak to valley waviness, but much depends on the periodicity. Per the definition:

waviness high, W[sub]t[/sub] - the peak-to-valley height of the modified profile from which roughness and part form have been removed by filtering, smoothing, or other means.

so, if B-C is the peak to valley roughness, then A-B would be the closest to what the profile would look like with the roughness removed, assuming that the part form has been removed as well.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[Belanger]

Pmarc, I should add that the picture didn't show the letters A, B, C (at least for me). That's why I wrote that "I will presume that A is the bottom blue line, B is the middle red line, and C is the top line."

If the letter do show up for you and I mistakenly assumed the wrong sequence, then that would be the source of the confusion!

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[IRstuff]

A is the top line, B is the middle, and C is the bottom.

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

 

[pmarc]

J-P,
There is no confusion. Based on you answer I assumed that the letters did not show up to you, and that is why I used terms "middle" and "bottom" line in order not to muddy further discussion.

But coming back to the main topic and bxbzq's graphics, I would say it is rather unlikely to assess form of surface and its roughness based on one picture. So if what we see is real picture, we can look at different surface characteristics depending on filtering frequency and stylus tip used, but not at all of them at once. In other words, if this plot is coming from flatness check, we probably can say that A-C is actual flatness error but do not see roughness parameters well and conclusions like "roughness is B-C distance" may be wrong.

 

[SeasonLee]

By definition of flatness in standard Y14.5-2009, roughness (surface texture) and flatness are independent, why we need to check roughness (and waviness) when measuring flatness?

Besides, the flatness tolerance is specified in millimeters and surface texture is specified in micrometers. Therefore, these are different unit and surface texture should not applicable to flatness control.

I'm just wondering, please let me know if I am wrong.

Season

 

[bxbzq]

Been away for a while.

Season,
As far as I know, there are applications where the roughness and form tolerances are at same precision level. Critical components in servo valve, for example, have form tolerances at micrometers while roughness is about Ra1 micrometer or less. In cases like this, form tolerance measurement methods need to be clearly defined.

 

[dingy2]

SeasonLee:
I agree with your thoughts. We check roughness with a profilometer and flatness with a indicator.

Dave D.

http://www.qmsi.ca

 

[Belanger]

Even if the flatness tolerance were down to a number such as 0.000001 (equivalent to 1 micron), it wouldn't be the same as roughness because the roughness of 1 micron would be measured over a cutoff length. This means that the part could be smooth as glass yet still be have a concavity of 2 mm, for instance.

John-Paul Belanger
Certified Sr. GD&T Professional
Geometric Learning Systems

 

[dgallup]

You can use a profilometer for flatness measurements, we do it all the time. It's possible to blow up the vertical scale by any value you want to see fine detail in the profile without any filtering. We have many lapped parts with flatness requirements in the micron range.

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

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.

 

[bxbzq]

OK. Different inspection tools give you different flatness results.

Look at the sketch attached in my post at 24 Aug 13 11:52,
By flatness definition, the flatness error is distance A-C. Yes or no?
Using indicator, the flatness error is close to A-B, right?
Using profilometer(this is something I'm not familiar), the flatness error is close to what, A-C?

 

[CheckerHater]

Using profilometer(this is something I'm not familiar), the flatness error is close to what, A-C?

It depends on your sampling length.
The idea is to keep sampling length short enough to keep profilometer reading close to B-C.
(You can find some additional info here:

http://www.bcmac.com/pdf_files/surface finish 101.pdf)

That allows to filter (separate) your surface texture from your flatness.

 

[dgallup]

Using the profilometer will give you a flatness measurement of approximately A-C, depending on your probe tip geometry. You don't use any cutoff length at all for the flatness, there is no filtering.

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

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.

 

[bxbzq]

Thank you guys.